阅读说明：本技术 一种针对水轮机活动导叶开口不匀的故障预警方法 (Fault early warning method for unevenness of opening of movable guide vane of water turbine ) 是由 颜天成 李太斌 张冲 延帅 吴家乐 姜东� 杜俊邑 于 2020-04-26 设计创作，主要内容包括：本发明公开了一种针对水轮机活动导叶开口不匀的故障预警方法,选取任意时间段内水轮机活动导叶开口不匀的相关监测量,组合神经网络算法和遗传算法建设智能算法模块,将相关监测量导入智能算法模块,得该时间段内水轮机开口不匀的最优值,将最优值作为评估水轮机活动导叶开口不匀故障的指标量,通过监测报警该指标量的变化趋势,实现对水轮机活动导叶开口不匀的故障预警。本发明对水轮机的相关监测量采用智能算法模块,对数据进行特殊的处理,从数据群体中挑选最优值,一方面能达到数据清洗的效果,另一方面从一组数据中提取到最能表达设备状态的特征值,这样既解决了数据由于信号干扰、传感器精度的误报问题,又能通过监测最优值实现故障预警。(The invention discloses a fault early warning method for the unevenness of the movable guide vane opening of a water turbine, which comprises the steps of selecting related monitoring quantity of the unevenness of the movable guide vane opening of the water turbine in any time period, combining a neural network algorithm and a genetic algorithm to construct an intelligent algorithm module, leading the related monitoring quantity into the intelligent algorithm module to obtain the optimal value of the unevenness of the opening of the water turbine in the time period, using the optimal value as an index quantity for evaluating the unevenness fault of the movable guide vane opening of the water turbine, and realizing the fault early warning of the unevenness of the movable guide vane opening of the water turbine by monitoring and alarming the change trend of the index quantity. The invention adopts an intelligent algorithm module for the related monitoring quantity of the water turbine, specially processes data, and selects an optimal value from a data group, so that on one hand, the effect of cleaning the data can be achieved, and on the other hand, a characteristic value which can best express the equipment state is extracted from a group of data, thereby solving the problems of signal interference and misinformation of the sensor precision of the data, and realizing fault early warning by monitoring the optimal value.)

1. The fault early warning method for the unevenness of the openings of the movable guide vanes of the water turbine is characterized by comprising the following steps of: selecting related monitoring quantity of the unevenness of the opening of the movable guide vane of the water turbine in any time period, combining a neural network algorithm and a genetic algorithm to construct an intelligent algorithm module, introducing the related monitoring quantity into the intelligent algorithm module to obtain an optimal value of the unevenness of the opening of the water turbine in the time period, taking the optimal value as an index quantity for evaluating the unevenness of the opening of the movable guide vane of the water turbine, and realizing the early warning of the unevenness of the opening of the movable guide vane of the water turbine by monitoring and alarming the variation trend of the index quantity.

2. The fault pre-warning method for the unevenness of the openings of the movable guide vanes of the water turbine as claimed in claim 1, wherein: the related monitoring quantities comprise monitoring quantities of top cover vibration, water guide swing, volute pressure pulsation, draft tube pressure pulsation, guide vane opening and a water head of the water turbine, and the monitoring quantities of the top cover vibration, the water guide swing, the volute pressure pulsation, the draft tube pressure pulsation, the guide vane opening and the water head are subjected to data normalization processing and then are led into the intelligent algorithm module.

Technical Field

The invention relates to the technical field of water turbine application, in particular to a fault early warning method for uneven openings of movable guide vanes of a water turbine.

Background

With the development of artificial intelligence, the monitoring of each part of equipment of the water turbine of each large hydropower plant in China at present still stays in the monitoring stage, the state of the equipment is judged through the manual analysis of real-time data, the analysis efficiency is low, and the misjudgment rate is high. Real-time monitoring may not be able to capture for aging of devices and gradual failure of devices. Due to the jump of a measuring point caused by interference factors such as sensor precision and the like, frequent false alarm and the like may occur on the degradation monitoring of equipment.

In order to break through the traditional data analysis method, a set of modules capable of realizing intelligent analysis, intelligent fault monitoring and intelligent fault prediction are required to be developed aiming at the uneven opening faults of the guide vane of the water turbine, so that intelligent early warning of the uneven opening faults of the guide vane of the water turbine is realized.

Disclosure of Invention

The invention aims to solve the problems and provide a fault early warning method for the uneven opening of the movable guide vane of the water turbine.

In order to achieve the purpose, the invention provides a fault early warning method for the unevenness of the movable guide vane opening of the water turbine.

The invention has the beneficial effects that:

the invention relates to a fault early warning method for uneven openings of movable guide vanes of a water turbine, which adopts an intelligent algorithm module for related monitoring quantity of the water turbine, performs special processing on each data individual in a data group, and selects an optimal value from the data group, so that the effect of cleaning the data can be achieved, and a characteristic value which can best express the state of equipment is extracted from a group of data, thereby solving the problems of signal interference and false alarm of sensor precision of the data, and realizing fault early warning by monitoring the optimal value.

Drawings

The accompanying drawings, which are included to provide a further understanding of the disclosure and are incorporated in and constitute a part of this specification, illustrate embodiments of the disclosure and together with the description serve to explain the disclosure without limiting the disclosure. In the drawings:

FIG. 1 is a flow chart of a fault early warning method for the opening unevenness of the movable guide vanes of the water turbine according to the invention;

FIG. 2 is a diagram of a BP neural network;

FIG. 3 is a flow chart of a portion of the genetic algorithm of the present invention;

FIG. 4 is a comparison graph of waveform frequency spectrum when the movable guide vane of the water turbine is normal and abnormal;

FIG. 5 is a graph of indicators of variable guide vane opening uniformity.

Detailed Description

The following detailed description of specific embodiments of the present disclosure is provided in connection with the accompanying drawings. It should be understood that the detailed description and specific examples, while indicating the present disclosure, are given by way of illustration and explanation only, not limitation.

The invention relates to a fault early warning method for the unevenness of the opening of a movable guide vane of a water turbine, which selects related monitoring quantities of the unevenness of the opening of the movable guide vane of the water turbine in any time period, wherein the related monitoring quantities comprise monitoring quantities of top cover vibration, water guide swing degree, volute pressure pulsation, draft tube pressure pulsation, guide vane opening and water head of the water turbine, and data normalization processing is carried out on the monitoring quantities of the top cover vibration, the water guide swing degree, the volute pressure pulsation, the draft tube pressure pulsation, the guide vane opening and the water head.

And establishing an intelligent algorithm module by combining a neural network algorithm and a genetic algorithm, introducing the related monitoring quantity after normalization processing into the intelligent algorithm module to obtain an optimal value of the unevenness of the opening of the water turbine in the time period, taking the optimal value as an index quantity for evaluating the unevenness fault of the opening of the movable guide vane of the water turbine, and realizing the fault early warning of the unevenness of the opening of the movable guide vane of the water turbine by monitoring and alarming the change trend of the index quantity.

The algorithm in the intelligent algorithm module comprises a neural network algorithm part and a genetic algorithm part.

The development of neural network algorithms, particularly multilayer neural network technology, provides a new method for the research of various different prediction physical models. The multi-layer neural network can continuously learn new knowledge and can process complex nonlinear mapping, and the BP model adopted by the neural network algorithm part in the invention is the most mature and effective model. For a prediction method based on the BP neural network theory, on the basis of the BP neural network, the data preprocessing before prediction, the influence factor quantification processing and the correction of a small amount of data after prediction are combined, so that the prediction model can reach the ideal precision, as shown in FIG. 2.

Genetic algorithms begin with a population representing a potential solution set to the problem, and a population is composed of a certain number of individuals that are genetically encoded. Each individual is actually a chromosome-bearing entity. After the initial generation population is generated, according to the principle of survival and the principle of excellence and disadvantage of fittest, generation-by-generation evolution generates better and better approximate solutions, in each generation, individuals are selected according to the fitness of the individuals in the problem domain, and combination crossing and variation are performed by means of genetic operators of natural genetics to generate a population representing a new solution set. This process will cause the population of the next generation, like natural evolution, to be more adaptive to the environment than the previous generation, and the optimal individuals in the population of the last generation, after decoding, can be used as a near-optimal solution to the problem, as shown in fig. 3.